The subduction dichotomy of Plate Tectonics
         (... "Ridge-push - slab-pull" ..  and other peregrinations of plates ... )

Fig.1.  The chicken-or-the-egg dichotomy.  Which is first, ..the lava or the slag?  Plate Tectonics says the slag is first....  (sort of.)    "What gets subduction going is an interesting problem."

 

(The chicken or the egg.)  Problem?  Well, ..if it's a problem, particularly in the face of such an onerous commandment, maybe we need to consider that it doesn't get going at all.  Maybe the answer has nothing to do with chickens or eggs, and lies elsewhere.

What Plate Tectonics is saying here in its hat-magic plate-spinning, chicken-or-the-egg way, is that the slag being extruded from the volcano creeps down the slope, pushes everything ahead of it ("ridge-push"), all the time cooling down, till it meets resistance in the form of a continental buttress.  Any earthy stuff ('continental crust') there might be getting carried along (which is relatively light), gets plastered on the the side of the buttress as folds which buckle up (like the Himalayas (??a bit of slag running down a hill slope??), whilst the slag you see in the figure gets forced down into the mantle by this buttress.  On getting forced down, the slag converts to its denser equivalent eclogite and falls into the mantle, whereupon the key is turned, the ignition sparked, and the subduction engine cranks into action and Plate Tectonis purrs happily away.  The falling slag ('slab") pulls the ocean floor down with it ("slab-pull"), ..everything, ... the whole ocean floor right back to the spreading ridge, thus depressurising the mantle underneath the ridge and causing the extrusion of the lava you see in the figure.   That is, the slag creeping down the slope in the figure is finally responsible for the extrusion of the magma (as well as pushing up the Himalayas).

:-)

Such a panoply!  But there is something else that has to happen before this occurs:-  the offsets on the spreading ridge before the Big "ca-Chunk"  from brittle failure to ductile flow; ... the reverse gear change that switched global tectonics from ridge offsetting in 'this' direction to ductile flow in 'that' direction, that caused the ocean floors to open.

It is hardly necessary to advance consideration of Plate Tectonics beyond this point of nonsense, because you and I know, even if intuitively (/unscientifically) experienced according to all reasonable sense of proportion, that this cannot be true, ...that it is clearly a case of magma first, slag second, and not, as Plate Tectonics says, the other way round.  I mean, ..how cold do rocks have to get before they can subduct?  And how can they be colder in a subduction zone hundreds of kilometres deep into the mantle  ("the cold subducting slab") than when they're traversing the ocean floor before getting there?  And why do they have to travel half way around the world anyway? .... why can't they just subduct right there, ..off the slope of the volcano?  Can't mantle subduct without the help of a continent to push it down?  Isn't mantle heavy /dense enough already?  And isn't that why it's already down there anyway, underneath the crust - because it is dense?  Surely it is pretty cold on the slope of an icelandic volcano (Fig.1), much colder than hundreds of kilometres into the mantle, where it is getting hot enough to convert the slag to eclogite.
 
 

Fig.2  Pot-on-the-stove analogy for Plate Tectonics.   Not realistic, since the heating element is below the liquid (mantle), whereas in reality the heat is in the liquid, which is a solid, and rises as a solid, ..and as a solid creates problems for the whole notion of 'convection'.  More about 'soup'. (Image courtesy of John Wiley and sons).

However subduction is indeed the foundation of Plate Tectonics so it needs to be addressed.  Gravitational correction ('cooling' as  'ridge-push - slab-pull')  is the equilibrium response to thermal rise.  So in emphasising subduction as the driver (aided by ridge-push) (whether silly or not), Plate Tectonics is essentially emphasising gravitational correction to the impetus of mantle rise as the driver, and thus effectively aligning itself in a very substantial way with Earth Expansion when it comes to the physical manifestion of deformation of the crust. Couple this with Plate Tectonics increasing recognition of  'overriding' rather than subduction, and 'flat subduction' rather than normal subduction, and there is even closer affinity between the two dynamical systems.

The problem for Plate Tectonics in claiming legitimacy to this alignment however is that all three (flat subduction, overriding, and gravitational collapse) are completely at odds with its central plank - cycling convection, and the assumption that the Earth cannot get bigger, for those three elements (gravitational correction, overriding, and flat subduction) are direct, demonstrable indications that the Earth is indeed getting bigger.   However in acknowledging this cost to itself, Plate Tectonics does appear to be making at least some progress towards Earth expansion.

Fig.3. Laki Volcano on Iceland.  Bags packed and tickets in hand this chunk of lithosphere would prefer the destination of a subduction zone, but 'overriding' is on the ticket, the closest being the American one for the plate on the right, and the further one being the Asian one, ...aaAAaaall the way across Europe for the plate on the left.  In Earth expansion it is of course nothing of the sort.  Everything about the Earth's dynamics is all happening RIGHT THERE (photo: top of the page) ...'everything' being the making of those hills (and their erosion) (Nice pics from photobucket.)

"..Magma first, slag second, ..that's it.
If subduction were the driver it would be the other way round.